Loading...

Russian scientists have developed a trained neural network of plastic

Memristor

The creation of new systems with artificial intelligence is still one of the most promising areas of modern science. Joint efforts of Russian physicists at the Kurchatov Institute, Moscow Physicotechnical Institute, St. Petersburg University, Moscow University and their Italian colleagues from Parma University created a neural network based on polymeric memristors.

For information, memristor and resistor are analogs. The main difference between a memristor and a “colleague” is that it can change its properties depending on the magnitude of the external signal. In this they are surprisingly reminiscent of synapses – the connections between two neurons of the brain, which also flexibly change the efficiency of signal transmission.

As the material for the manufacture of memristors, the polymer was chosen as polyaniline. Then the scientists began to learn the artificial neural network created by them. To do this, pulses were sent to its input in a random sequence. In case of an incorrect answer, a special corrective impulse was immediately sent. The process of learning the network consisted in a repetition – at least 15 times – of such operations, after which the system was tuned to the “right mode” and was able to perform basic logical operations.

Researchers are only at the beginning of the journey. Ahead of them it is necessary to minimize the size of memristors to the size of 10 nanometers, to increase the speed of their reaction to incoming signals. Only after this it will be possible to talk about the practical application of the development.

Leave a Reply

Your email address will not be published. Required fields are marked *